Cesarean births are much more expensive than vaginal births and are associated with substantially increased maternal morbidity. Cesareans are needed to reduce fetal and maternal risk in selected pregnancies, but there is general agreement that the U.S. cesarean rate, which still exceeds 20%, is higher than is necessary. A target of 15% was set in Healthy People 2000, and, because there is no chance that it will be achieved next year, has been left unchanged in the draft version of Healthy People 2010. A major obstacle to achieving this goal is the paucity of evidence showing that, in the context of U.S. obstetric practice, a low cesarean rate is just as safe for the fetus/infant as a high rate. We propose a novel strategy to obtain such evidence. The primary objectives of this proposed study are to evaluate the effectiveness of frequent use of cesarean section in reducing perinatal mortality, and, if it is effective, to estimate the cost per infant/fetus saved as measured by hospital charges. Comparing birth outcomes between cesarean and vaginal delivery is inherently biased, because patients delivered by the two methods have different risks. To avoid this trap, the proposed study addresses the question at the physician level. Physicians are grouped by their cesarean section rates and birth outcomes are compared between physician groups. Using linked data from New Jersey's electronic birth certificate and maternal hospital discharge data, differences in risk profile between populations served by these physician groups are taken into account. The purpose is to determine whether, after adjusting patient risk differences, physicians with high cesarean section rates have better (or worse) birth outcomes than those with lower rates. For this study, data from New Jersey birth certificates, fetal death certificates, infant death certificates, and maternal and infant hospital discharge records will be linked for the years 1996 and 1997 (estimated 220,000 births). Physicians with more than 100 deliveries over the two- year period will be classified into three groups based on their cesarean section rates: low (less than or equal to l5%), medium (15%-25%), and high (greater than 25%). To minimize bias in grouping, perinatologists will be identified and excluded. Based on preliminary analysis, 12%, 44%, and 44% of physicians, respectively, are anticipated to be classified to the three groups. Primary outcomes are perinatal mortality including fetal death (more than 28 weeks of gestation), early neonatal death (less than 7 days), neonatal death (less than 28 days), and health resource use including length of hospital stay and charges. In the event that high-rate physicians have fewer perinatal deaths, the extra hospital charges incurred per fetus saved will be estimated. Hospital charges will be used as a surrogate for direct hospital costs since the latter are not available.